salinity effects on seedling growth and yield components ... · salinity effects on seedling growth...

Salinity Effects on Seedling Growth and Yield Components of Rice

Linghe Zeng* and Michael C. Shannon

ABSTRACT reduction were observed (Scardaci et al., 1996; Shannonet al., 1998).Flood irrigation practices that are commonly used in California

Rice is rated as an especially salt-sensitive crop (Maasduring the early stages of rice (Oryza sativa L.) establishment mayand Hoffman, 1977; Shannon et al., 1998). The responsecontribute to salinity damage and eventually decrease yield. Knowl-of rice to salinity varies with growth stage. In the mostedge of salinity effects on rice seedling growth and yield components

would improve management practices in fields and increase our under- commonly cultivated rice cultivars, young seedlingsstanding of salt tolerance mechanisms in rice. Salinity sensitivity of were very sensitive to salinity (Pearson and Bernstein,rice was studied to determine salinity effects on seedlings and yield 1959; Kaddah, 1963; Flowers and Yeo, 1981; Heenan etcomponents. Plants of rice cultivar M-202 were grown in a greenhouse al., 1988; Lutts et al., 1995). Yield components related toin sand and irrigated with nutrient solutions of control and treatments final grain yield were also severely affected by salinity.amended with NaCl and CaCl2 (2:1 molar concentration) at 1.9, 3.4, Panicle length, spikelet number per panicle, and grain4.5, 6.1, 7.9, and 11.5 dS m21 electrical conductivity. Shoot dry weights yield were significantly reduced after salt treatmentsof seedlings were measured at five harvests in the first month after

(Sajjad, 1984; Heenan et al., 1988; Cui et al., 1995; Kha-seeding. Seedling growth was significantly reduced by salinity at thetun et al., 1995). Salinity also delayed the emergence oflowest salinity treatment, 1.9 dS m21. At 1.9 and 3.4 dS m21, significantpanicle and flowering (Khatun et al., 1995) and de-reduction of seedling growth occurred at longer cumulative thermalcreased seed set through reduced pollen viability (Kha-time than at higher salt levels. Seedling survival was significantlytun and Flowers, 1995; Khatun et al., 1995). In contrast,reduced when salinity was 3.40 dS m21 and higher. Highly significant

linear responses of grain weight per plant, grain weight per panicle, rice was more salt tolerant at germination than at otherspikelet number per panicle, and tiller number per plant to salinity stages (Narale et al., 1969; Heenan et al., 1988; Khanwere observed. There was a common lowest salt level for fertility and et al., 1997). Seed germination was not significantly af-pollen germination beyond which they were significantly reduced by fected up to 16.3 dS m21, but was severely inhibitedsalinity. Harvest index was significantly decreased when salinity was when salinity increased to 22 dS m21 (Heenan et al.,at 3.40 dS m21 and higher. Tiller number per plant and spikelet number 1988). The suppression of germination at high salt levelsper panicle contributed the most variation in grain weight per plant

might be mainly due to osmotic stress (Narale et al.,under salinity. Reductions in seedling survival, tiller number per plant,1969; Heenan et al., 1988).and spikelet number per panicle were the major causes of yield loss

Although there are extensive studies of salinity effectsin M-202 under salinity. The compensation between spikelets andon rice, our understanding of the quantitative effectsother yield components was confounded with salinity effects, but wasof salinity on rice and critical thresholds of responses,believed to be minor relative to the reduction of spikelets due to

salinity and, therefore, not sufficient to offset yield loss even at moder- especially with respect to modern, commonly used culti-ate salt levels. vars, is still limited. The determination of salinity-sensi-

tivity parameters, for example, thresholds of salinityeffects on rice seedling growth and yield componentsand the interrelationships among yield components un-Salinity is one of the major obstacles to increasingder salinity stress, will help to develop better manage-production in rice growing areas worldwide. In re-ment practices for growing rice under salinity and im-cent years, the use of recirculating water systems in riceprove our understanding of the mechanisms of saltproduction and the requirement for water holding intolerance in crops. The objectives of this study were tothe fields without recirculating water systems have in-determine the effects of salt levels and stress durationscreased in areas of the USA because of environmentalon seedling growth under controlled conditions, to ana-concerns related to the problems caused by drainagelyse the salinity sensitivity of different yield components,into receiving rivers after pesticide applications (Scar-to identify the most important yield component(s) caus-daci et al., 1996). The use of recirculating water systemsing the reduction of grain yield, and to determine thecoupled with regulations for water holding have reducedpossible compensations between yield components thatthe hazard to the environment, but have also increasedcould completely or partially offset the yield loss undersalinity levels in the standing water of rice fields. In amoderate salinity.rice field survey in California, the electrical conductivity

(EC) of field water during the water-holding periodMATERIALS AND METHODSafter pesticide application ranged from 0.5 to 3.5 dS

m21, with the lowest EC values in the top basin and the Plant Materialshighest in the bottom basin (Scardaci et al., 1996). In

A medium-size grain, early maturity, and semidwarf ricethese salt-affected rice growing areas, substantial losscultivar, M-202, was used. The seed was obtained from C.in seedling establishment, leaf chlorosis, and final yield Johnson (California Cooperative Rice Research Foundation,Biggs, CA). Like most other common cultivars in California,M-202 has been rated as very sensitive to salinity (Shannon

L. Zeng and M.C. Shannon, U.S. Salinity Laboratory, USDA-ARS, et al., 1998).PWA, Riverside, CA 92507. Received 23 Oct. 1998. *Correspondingauthor ([email protected]).

Abbreviations: DAP, days after planting; EC, electrical conductivity;PI, panicle initiation.Published in Crop Sci. 40:996–1003 (2000).

996

Published August 21, 2014

kailey.harahan

Typewritten Text

1613

ZENG & SHANNON: SALINITY EFFECTS ON RICE SEEDLINGS AND YIELD COMPONENTS 997

which were randomly chosen from each replicate. The seedlingPlant Culturesurvival rate was calculated as the percentage of live seedlings

Two experiments were conducted in the greenhouse at Riv- from germinated plants. At each harvest at seedling stage,erside, CA (33858924″ N, 117819912″ W) between May and 15 seedlings of each replicate were randomly sampled fromNovember 1997. The plants were cultured in nutrient solution surviving plants. After roots were removed, shoots of seedlings(Yoshida et al., 1971) in sand tanks (122 by 61 by 46 cm deep) were dried in a forced-air oven (708C) for 1 wk. The dryfilled with white silica sand (#12, Cisco Inc., CA)1 with an weights of seedlings were measured using an electronic bal-average bulk density of 1.4 g cm23. Nutrient solution pH was ance with milligram precision.maintained between 5.5 and 6.5 by adding H2SO4 twice a week. Main culms of all surviving plants were tagged before tiller-Nutrient solutions were changed once in the middle of the ing. After harvest, main culms and panicles on main culmsgrowing season. Nutrition status of plants was monitored by were separated from the other culms and panicles. Plantsvisual observations. Irrigation solutions were prepared in were bagged individually after roots were removed. Shoot dry1100-L reservoirs and pumped to provide irrigation to the weights of main culms and all tillers were measured after ovensand tanks. Overflow irrigation was returned through drainage drying at 708C for 1 wk. Panicles on main culms were countedby gravity to the reservoirs. Each reservoir provided irrigation and weighed. Panicles on tillers were also weighed. The follow-to six sand tanks (replicates) three times daily for 30 min per ing yield components were determined: primary branches perirrigation cycle. Rice seed was sterilized in HgCl2 (0.5g L21) panicle, panicle length, spikelet number per panicle, fertility,for 2 min, rinsed with formaldehyde (16 mL L21) and methanol 1000-kernel weight, and grain dry weight per panicle. Among(4 mL L21) for 15 min, and then rinsed and soaked in distilled them, fertility was defined as the percentage of filled spikeletswater for 24 to 48 h. Seeds were planted in nine rows per relative to the total number of spikelets per panicle. Tillertank. The rows were spaced 12 cm apart with 15 seeds per number per plant and grain dry weight per plant were alsorow. Sowing depth was ,1 cm. Water depth was controlled measured. Tiller number per plant was determined on allat 1 to 2 cm during the first week and at 6 to 8 cm thereafter. tillers with emerged heads. All matured panicles (i.e., kernelsHourly mean temperatures were integrated across the 24 h were too hard to be dented by the thumbnail) were hand-and summed to give cumulative thermal time (8C d, Logan threshed and weighed for grain dry weight per plant. Theand Boyland, 1983). Air temperature ranged from 25 to 338C immature panicles were not weighed. Harvest index was calcu-during the day and 18 to 238C during the night. Humidity lated as grain dry weight per plant divided by the totalranged from 40 to 85%. Light averaged 1050 mmol m22 s21

aboveground biomass, which was the sum of grain dry weightwith a minimum of 200 and a maximum 1400 mmol m22 s21 at per plant and shoot dry weight per plant.noon. A mixture of NaCl and CaCl2 (2:1 molar concentration) Most yield components were only measured from the pani-were added to the nutrient solutions 5 d after seeding. The cles on main culms because of the consideration of nonuniformstress remained continuous till harvest. Electrical conductivi- maturity among tillers and plants and the labor involved inties of nutrient solutions were measured with an EC meter making the measurements. Some information on other pani-on alternate days. A control (i.e., nutrient solution without cles such as those on young tillers was sacrificed by thisNaCl and CaCl2) was prepared. A randomized block design method; however, the results obtained from these measure-was used with five replicates for each salt treatment in the ments still represent, in a general sense, the salinity effectsfirst experiment and six replicates in the second experiment. on all yield components.

For measurements of pollen viability, two plants from eachPlant Harvest replicate were randomly chosen. About one hundred pollengrains in each of two plants were scored for both pollen stain-In the first experiment, seeds were planted 22 May 1997.ability and in vitro germination. Pollen grains were stainedThe survived seedlings were harvested at cumulative thermalwith 0.9% (w/v) thiazolyl blue in 54% (w/v) sucrose (Khatuntimes of 276 [i.e., 11 d after planting (DAP)], 349 (i.e., 14and Flowers, 1995). Fresh pollen grains were collected onDAP), 420 (i.e., 17 DAP), 495 (i.e., 20 DAP), and 5668C dmicroscope slides in a drop of fresh stain solution and scored(i.e., 23 DAP), assuming a base temperature of 08C. Seedlingfor stainability after 5 to 10 min. For in vitro pollen germina-dry mass accumulation was measured at these time intervals.

Plants were not grown to maturity in the first experiment. tion, pollen grains were germinated in agar medium, as de-In the second experiment, seeds were planted on 17 July scribed by Brewbaker and Kwack (1963), with the following

1997. The survived seedlings were harvested at cumulative components: 1.6 3 1023 M H3BO3, 1.3 3 1023 M Ca(NO3)2,thermal times of 289 (i.e., 12 DAP), 365 (i.e., 15 DAP), 469 8.1 3 1024 M MgSO4, 1 3 1023 M KNO3, 2.7 3 1025 M EDTA,(i.e., 19 DAP), 550 (i.e., 22 DAP), and 6228C d (i.e., 25 DAP), 20% (w/v) sucrose, and 0.5% (w/v) agar, at pH 5.5. Freshassuming a base temperature of 08C, for measurements of pollen grains were collected on agar in petri dishes immedi-seedling dry mass accumulation. The seedling survival rate ately after anthesis and cultured at 378C for 20 min beforewas only measured in the second experiment at 6228C d. or observation under a light microscope at 1003 magnification.25 DAP. The dead plants were removed and the remaining The yield measurements of each salt treatment were com-plants in each sand tank were thinned to 12 cm between rows pared and grouped with Tukey’s multiple comparisons (Ott,and 10 cm between plants. Plants were harvested in November 1993). Regression relating seedling survival rates to salinity1997 for yield component analysis. Plants at different salt levels was analysed with a quadratic regression model. Alllevels were harvested sequentially when seed on primary tillers statistical analyses were carried out with the SAS system (SASmatured (i.e., the kernels were yellow and could no longer to Institute, 1994).be dented by the thumbnail).

Interrelationships among Yield ComponentsMethods of Measurements for Salinity Sensitivityat Different Growth Stages The relative importance of yield components to grain yield

(i.e., grain weight per plant, GWL) was evaluated based onSeedling survival rate was measured in three rows of seed-the equation of GWL as the product of tiller number per plantlings [germinated from 15 (no. row21) 3 3 (rows) 5 45 seeds],(TP) and grain weight per panicle (GWP). Since GWP is thedirect function of spikelet number per panicle (SP), fertility1 Mention of trade names or commercial products in this article is(FT), and 1000-kernel weight (TW), the equation can be ex-solely for the purpose of providing specific information and does not

imply recommendation or endorsement by the USDA. pressed as

998 CROP SCIENCE, VOL. 40, JULY–AUGUST 2000

Table 1. Means and standard errors of shoot dry weights of survived rice seedlings harvested at different cumulative thermal times afterplanting at different salt levels (Exp. 1).

Thermal time (8C d)†

276 349 420 495 566Salt (11 DAP) (14 DAP) (17 DAP) (20 DAP) (23 DAP)

dS m21 mg plant21

Control 15.8 6 0.8a‡ 20.2 6 0.8a 32.2 6 1.2a 63.5 6 3.0a 123.6 6 7.1a1.9 13.9 6 0.7ab 17.5 6 0.9ab 28.1 6 1.5ab 46.3 6 3.1b 83.8 6 5.0b3.2 13.5 6 0.7ab 18.5 6 0.8ab 23.5 6 1.5bc 46.8 6 3.2b 70.2 6 4.7bc4.6 12.9 6 0.6b 15.4 6 0.9bc 20.3 6 1.6cd 39.8 6 2.9bc 62.3 6 4.3cd6.2 11.4 6 0.5b 14.1 6 0.9c 22.9 6 1.2bc 36.8 6 2.6bc 52.2 6 3.3cd8.1 11.5 6 0.6b 15.5 6 0.7bc 19.9 6 0.9cd 34.2 6 2.1c 54.5 6 3.3cd11.7 12.1 6 0.4b 13.3 6 0.5c 17.2 6 0.8d 30.9 6 1.5c 45.3 6 2.4d

† Cumulative thermal time was summed from hourly mean temperatures that were integrated across the 24 h.‡ Means within columns followed by the same letter are not different at P 5 0.05 according to Tukey multiple comparisons.

GWL 5 TP 3 SP 3 FT 3 TW of salinity treatments were averaged through the dura-tion of salinity stress, continuous stress through harvest,The relative importance of yield components to grain yieldto six salt levels: 1.9 (1.6–2.0), 3.4 (3.1–3.7), 4.5 (4.1–4.8),was analyzed using multiple linear regression. The significance6.1 (5.8–6.7), 7.9 (7.5–8.3), and 11.5 (10.5–12.1) dS m21.of yield components was determined by stepwise regression

analysis, that is, based on the order of their addition into the The EC of controls was 0.87 dS m21 in the first experi-multiple regression model. ment and 1.15 dS m21 in the second experiment.

Tiller number per plant depended on plant density (Counceet al., 1989; Wu et al., 1998), which was determined at vegeta-

Salinity Effects on Seedling Growthtive stages. Spikelet number per panicle was determined atpanicle initiation, fertility at booting and early anthesis, and The results of the first and second experiments1000-kernel weight at filling stage (Yoshida et al., 1981; Hoshi- showed that salinity caused a significant reduction inkawa, 1989). It is clear that these yield components are sequen-

seedling growth very soon after planting (Tables 1 andtially and successively formed in the order of tiller number2). In Exp. 1, significant (P , 0.05) growth reductionper plant, spikelet number per panicle, fertility, and 1000-occurred at 2768C d when salinity was 4.6 dS m21 andgrain weight. There was a compensatory relationship betweenhigher (Table 1). In Exp. 2, significant (P , 0.05) reduc-two successively formed yield components, with a certain level

of competition within a population. The potential value of a tion of seedling growth occurred at 2898C d when salinitynewly formed yield component decreased as the previously was 4.5 dS m21 and higher (Table 2). As the duration offormed yield component increased (Siband et al., 1999). A salinity stress increased, significant reduction in seedlingratio between two successively formed yield components, growth occurred at lower salt levels. In Exp. 1, a signifi-newly formed yield component/earlier-formed yield compo- cant reduction in seedling growth was first observed atnent, was used to study the interrelationships among these

2768C d when salinity was 4.6 dS m21. At 3.2 dS m21, ayield components under salinity. When density was fixed, instress period of 4208C d resulted in a significant reduc-populations of equal spaced plants, the changes of the ratiostion in seedling growth. At 1.9 dS m21, the lowest salinitywere expected to be a combination of salinity effects and atreatment, a stress period of 4958C d resulted in a signifi-possible compensation between the two successively formed

components. cant reduction in seedling growth compared with thecontrol (Table 1). In Exp. 2, a significant reduction inseedling growth was first observed at 2898C d whenRESULTSsalinity was 4.5 dS m21. After 3658C d, salinity treat-

Salinity Levels and Controls ments at 1.9 and 3.4 dS m21 resulted in a significantreduction in seedling growth compared with the controlIn the first experiment, the EC readings of salinity(Table 2).treatments were averaged through the duration of salin-

The percentage of seedling survival rate was signifi-ity stress to six salt levels: 1.9 (1.7–2.0), 3.2 (2.8–3.5),cantly (P , 0.05) reduced at 3.4 dS m21 and higher4.6 (4.2–4.9), 6.2 (5.6–6.7), 8.1 (7.4–8.5), and 11.7 (11.2–

12.1) dS m21. In the second experiment, the EC readings compared with the controls in Exp. 2 (Table 3). A qua-

Table 2. Means and standard errors of shoot dry weights of survived rice seedlings harvested at different cumulative thermal time afterplanting at different salt levels (Exp. 2).

Thermal time (8C d)†

289 365 469 550 622Salt (12 DAP) (15 DAP) (19 DAP) (22 DAP) (25 DAP)

dS m21 mg plant21

Control 12.6 6 0.6a‡ 20.8 6 0.8a 40.5 6 1.8a 123.9 6 6.7a 232.4 6 14.8a1.9 11.5 6 0.4ab 15.0 6 0.5b 24.7 6 1.4b 76.5 6 5.0b 161.4 6 12.2b3.4 11.1 6 0.4abc 14.2 6 0.6b 21.1 6 1.0b 67.7 6 4.5b 136.1 6 9.7bc4.5 10.8 6 0.4bc 13.5 6 0.5b 16.7 6 0.6c 45.8 6 3.6c 109.2 6 9.1c6.1 9.8 6 0.3cd 10.5 6 0.4c 12.5 6 0.5c 20.4 6 1.7d 42.4 6 4.8d7.9 8.9 6 0.3d 10.3 6 0.4c 12.6 6 0.6c 16.0 6 0.9d 24.5 6 2.5d11.5 8.7 6 0.3d 10.7 6 0.3c 12.7 6 0.5c 18.5 6 1.3d 25.2 6 2.4d

† Cumulative thermal time was summed from hourly mean temperatures which were integrated across the 24 h.‡ Means within columns followed by the same letter are not different at P 5 0.05 according to Tukey multiple comparisons.


Table 3. Means and standard errors of seedling survival rates, shoot dry weights of main stem, harvest index, and pollen viability atdifferent salt levels.

Salt levels (dS m21)

Control 1.9 3.4 4.5 6.1 7.9 11.5

%Seedling survival, %‡ 98.3 6 1.3a† 91.5 6 3.0ab 79.4 6 2.0bc 68.3 6 5.5cd 58.3 6 6.7de 56.0 6 1.8de 50.7 6 2.9eShoot weight g main culm21 4.11 1 0.11a 4.12 1 0.12a 3.64 1 0.11a 3.75 1 0.14a 2.89 1 0.13b 2.91 1 0.11b 2.20 1 0.09cHarvest index§ 0.49 1 0.01a 0.47 1 0.01a 0.41 1 0.01b 0.39 1 0.01bc 0.36 1 0.01c 0.29 1 0.01d 0.21 1 0.01ePollen stainability, % 76.9 6 2.5a 70.6 6 3.8a 65.1 6 3.0a 43.9 6 5.9b 28.2 6 3.7c 20.3 6 2.7cd 11.3 6 2.8dPollen germination, %¶ 51.8 6 5.7a 52.7 6 5.5a 49.8 6 4.1a 51.9 6 2.8a 29.2 6 5.0b 15.0 6 3.9bc 8.5 6 3.0c

† Means within rows followed by the same letter are not different at P 5 0.05 according to Tukey multiple comparisons.‡ Seedling survival was measured at 622 8C d. The survival rates are the percentages of live seedlings from germinated plants.§ Harvest index was calculated as the grain weight per plant divided by the total aboveground biomass per plant.¶ Pollen grains were stained in thiazolyl blue solutions (see Materials and Methods). Pollen stainability was calculated as the percentage of light red or

red pollen grains among the total observed. The germination percentage was calculated based on number of germinated pollens among the total observed.

dratic equation was determined to best predict the re- significant when compared with the control. Linear re-gressions of salinity on the yield components were highlysponse of seedling survival rates to salinity by stepwise

analysis (Fig. 1). The confidence intervals of seedling significant except for kernel weight (Fig. 2). When grainyield per plant was expressed as the product of tillersurvival rates were 81.1 to 74.4% at 3.4 dS m21 salt level

and 104.1 to 93.0% at control salt level. Note that the number per plant, spikelet number per panicle, fertility,and kernel weight, tiller number per plant, and spikeletseedling survival rates were almost linearly decreased

by salinities ,6.2 dS m21 (Fig. 1). Then the responses number per panicle contributed the most variation tograin yield, while fertility and kernel weight contributedbecame curvilinear between 7.9 and 11.5 dS m21.less, based on stepwise analysis (Table 5). The ratio ofspikelets per panicle/tillers per plant decreased with theSalinity Effects on Yield Componentsincrease of salinity (Table 6). The ratios of fertility/All yield components investigated, except 1000-ker-spikelets per panicle and kernel weight/spikelets pernel weight, were significantly (P , 0.05) reduced at 6.1panicle increased with the increase of salinity. The ratiodS m21 and higher compared with the controls (Tableof fertility/spikelets per panicle at 11.5 dS m21 was 1.94). None of the components was significantly (P , 0.05)times that of the control. The ratio of kernel weight/reduced at 1.9 dS m21. Spikelet number per panicle,spikelets per panicle at 11.5 dS m21 was 3.4 times thatgrain dry weight per panicle, and grain dry weight perof the control.plant were significantly (P , 0.05) reduced at 3.4 dS

m21 and higher. Tiller number per plant was significantly Salinity Effects on Other Relative Traits(P , 0.05) reduced at 4.5 dS m21 and higher. Primarybranches per panicle, panicle length, and fertility were Shoot dry weights of main culms were not significantly

(P , 0.05) reduced by salinity until EC was 6.1 dS m21not significantly reduced except at salinity levels of 6.1dS m21 and higher. Kernel weight was not significantly or higher (Table 3). Harvest indices were significantly

(P , 0.05) reduced by salinity at 3.4 dS m21 or higherreduced at any salt level tested. Instead, there was slightincrease of kernel weight at 11.5 dS m21, although not (Table 3). The pollen viability based on staining was

Fig. 1. Regression of salinity (measured by electrical conductivity [EC] of irrigation solutions) on seedling survival rate. Measurements of seedlingsurvival rates at 6228C d (25 DAP) from six replicates are represented by the symbol (3). The solid line is the quadratic equation: survivalrate (%) 5 111.36 2 11.83EC 1 0.57EC2 (F 5 76.21, df 5 2, 39). The dashed lines are upper and lower confidence levels (95%).


Table 4. Means and standard errors of yield components measures at different salt levels.

Salt Tiller Primary branch Panicle length Spikelet Fertility† Grain weight

dS m21 no. plant21 no. panicle21 cm no. panicle21 % g panicle21 g 100021 g plant21

Control 15.7 6 0.7a‡ 13.4 6 0.2a 20.8 6 0.2ab 176.3 6 4.7a 80.5 6 1.7a 4.1 6 0.1a 29.3 6 0.4ab 34.3 6 1.8a1.9 13.9 6 0.7ab 12.9 6 0.2a 21.2 6 0.2a 162.7 6 5.3a 76.9 6 2.0ab 4.0 6 0.1a 30.6 6 1.2ab 30.5 6 1.4ab3.4 13.1 6 0.9abc 11.5 6 0.3bc 18.5 6 0.7c 136.9 6 5.7b 74.1 6 2.0ab 3.0 6 0.2b 29.0 6 0.4ab 24.2 6 2.4bc4.5 12.2 6 0.9bcd 12.2 6 0.3ab 19.6 6 0.3bc 131.6 6 5.1b 72.8 6 2.6ab 2.6 6 0.1b 27.9 6 0.5ab 19.3 6 1.9c6.1 10.7 6 0.7cd 10.6 6 0.3c 18.2 6 0.4c 96.8 6 4.9c 70.1 6 2.9b 1.9 6 0.1c 27.0 6 0.6a 12.5 6 1.3d7.9 9.9 6 0.7d 10.4 6 0.3c 18.1 6 0.3c 93.2 6 3.7c 52.2 6 2.4c 1.4 6 0.1c 28.7 6 1.2ab 7.79 6 0.7de11.5 6.0 6 0.4e 8.7 6 0.3d 15.0 6 0.4d 53.9 6 3.5d 44.7 6 3.0c 0.8 6 0.1d 31.2 6 1.3b 2.52 6 0.2e

† Fertility was defined as the percentage of filled spikelets relative to the total number of spikelets per panicle.‡ Means within columns followed by the same letter were not different at P 5 0.05 according to Tukey multiple comparisons.

not significantly reduced except at 4.5 dS m21 and and 3.4 dS m21) significant reduction of seedling growthoccurred at higher cumulative thermal time than thosehigher, while the viability based on germination was not

significantly reduced except at 6.1 dS m21 and higher seedlings at higher salt levels (i.e., 4.5 dS m21 andhigher). This indicated that plants affected by salts atcompared with controls (Table 3). Note that there was

a steep decline in pollen stainability and germination low concentrations can tolerate salt stress for longerdurations before significant reduction of seedlingwhen salinity was above 3.4 and 4.5 dS m21, respectively

(Fig. 3). In contrast, a steep decline in fertility occurred growth occurs.Yield sink capacity is always one of the primary objec-when salinity was above 6.2 dS m21 (Fig. 3).

tives in crop breeding for increasing crop yield. Yieldsink capacity can be defined as the maximum size ofDISCUSSIONthe sink organs to be harvested (Kato and Takeda,

Reduction in seedling growth and loss of stand due 1996). The final grain yield can be described as theto salinity have been implicated as causative factors for product of the number of panicles per unit area andyield losses in California rice production (Scardaci et panicle weight. The number of panicles per unit areaal., 1996; Shannon et al., 1998). Our study indicated that depends on plant density and tillering ability of plants.salinity as low as 1.9 dS m21 can significantly reduce The low threshold value of salinity on rice M-202 seed-seedling shoot dry weight and that salinity at 3.4 dS m21 ling survival rate at the early seedling establishmentcan reduce seedling survival (plant stand) in M-202, a indicated a reduction in yield sink capacity under salinitycommon rice cultivar in California. The reductions in by a reduction in plant density. It was also known that,seedling survival rates and growth are major causes of under nonstressed conditions, there was a compensatorythe stand loss in salt-affected rice fields. The threshold relationship between plant density and tillering, whichvalue of salinity on seedling survival rate in M-202 is maintained panicle density within a certain range ofbetween 1.9 and 3.4 dS m21, although the exact value change in plant density (Counce et al., 1989; Hoshikawa,was difficult to determine because of the quadratic re- 1989; Counce and Wells, 1990; Gravois and Helms, 1992;sponse of this variable to salinity and the variability due Wu et al., 1998). The determination of this compensa-to the experimental conditions. The observed effects of tory relationship between plant density and tillering un-salinity on seedling growth were a function of both salt der salinity stress requires further studies at the popula-

tion level. However, the high sensitivity of tillering tolevel and time of exposure. At low salt levels (i.e., 1.9

Fig. 2. Regressions of salinity on relative yields based on yield components. Relative yield were calculated from observations for yield componentsdivided by the mean values of the controls. TW, 1000-kernel weight; PL, panicle length; PBP, primary branch per panicle; FT, fertility; TP,tiller number per plant; SP, spikelet number per panicle; GWP, grain weight per panicle; GWL, grain weight per plant; EC, electrical conductivity.


Table 6. Relationships between two successively formed yieldTable 5. Contributions of yield components to grain yield in riceunder salinity. Grain weight per plant was the product of tillers components under salinity stress.per plant (TI), spikelets per panicle (SP), fertility (FT), and

Salt (dS m21)1000-kernel weight (TW).Control 1.9 3.4 4.5 6.1 7.9 11.5Contribution to total

Yield components variation in yield (R2)† SP/TI† 12.3 11.7 10.7 10.8 9.3 9.5 9.2FT/SP 0.43 0.47 0.54 0.58 0.73 0.56 0.83TI 71.4 TW/SP 0.17 0.19 0.21 0.22 0.28 0.31 0.58SP 71.1

FT 38.0 † SP, spikelet number per panicle; TI, tiller number per plant; FT, fertility;TW 1.1 TW, 1000-kernel weight. The ratios of two successively formed yieldTI 1 SP 86.8 components were calculated from each observation of late-formed com-TI 1 SP 1 FT 89.1 ponent divided by each observation of early-formed component. TheTI 1 SP 1 FT 1 TW 89.9 order of the formation in yield components was assumed to be TI, SP,

FT, and TW.† The relationship between grain weight per plant and other yield compo-nents was determined by stepwise regression analysis. The yield compo-nents with highest R2 values were the first to add to the regression model. maximum yield. A compensatory relationship betweenThe next yield components were added to the model at P 5 0.05.

two successively formed yield components has been ob-served in maize (Zea mays L.). The newly formed yieldsalinity in M-202 indicated a limited tillering ability ofcomponent decreased when the earlier formed compo-rice plants under salinity stress. Yield sink capacity pernent increased (Siband et al., 1999). These phenomenaplant was defined as the product of tiller number perhave been observed in rice crops under nonstressedplant, spikelet number per panicle, fertility, and kernelconditions. Generally, the correlations between tillerweight (Kato and Takeda, 1996). In M-202, tiller numbernumber per plant and spikelet number per plant, andper plant and spikelet number per panicle were the mostbetween spikelet number per plant and fertility andsalinity-sensitive yield components, showing highly sig-kernel weight are negative (Counce and Wells, 1990;nificant linear responses to salinity (Fig. 2). These twoKato and Takeda, 1996). In wheat (Triticum aestivumyield components contributed the most variation toL.), a complete or partial compensation between spike-grain yield under salinity stress, based on stepwise analy-lets and kernels was observed which resulted in a nonsig-sis (Table 5). It was concluded that the reductions innificant reduction in final kernel number under moder-seedling survival rate, tiller number per plant, and spike-ate salinity (Maas and Grieve, 1990; Grieve et al., 1992;let number per panicle were the major causes of yieldMaas et al., 1996). The removal of some spikelets in-loss in M-202 under salinity.creased fertility on remaining spikelets in rice (Matsu-The yield components of tiller number per plant,shima, 1970). In the analysis of interrelationshipsspikelet number per plant, fertility, and kernel weightbetween these yield components under salinity stress,were believed to have their own critical developmentthe compensation was difficult to observe because ofperiods that can affect the final grain yield. Differentthe confounding effects with salinity. For example, asrelationships between these sequentially and succes-

sively formed yield components allow the calculation of the earlier-formed yield component, tiller number per

Fig. 3. Means of pollen viability and fertility.


plant, decreased, the expected compensation was an rapidly after their initiation (Hoshikawa, 1989). In riceM-202, the number of primary ranchi-branches and pan-increase in spikelet number per panicle. However, salin-

ity stress at PI reduced spikelet number per panicle. In icle length were significantly reduced by salinity at 6.1dS m21 or above, but not at salinities below that level,M-202, the ratio of spikelets per panicle/tillers per plant

decreased with the increase of salinity (Table 6). This while spikelet number per panicle was significantly re-duced at 3.4 or above (Table 4). The higher salinityindicated that the compensation between tiller number

per plant and spikelet number per panicle was limited threshold values of primary branches and panicle lengthcompared with spikelet number per panicle suggest thatand that the effect of salinity on spikelet number per

panicle was more pronounced than that on tiller number the degeneration of primary branches on panicle andthe reduction in panicle length were not major causesper plant. In contrast, the ratio of fertility/spikelets per

panicle increased almost two times, and the ratio of for the reduction of spikelets at moderate salinity. Cuiet al. (1995) observed the salinity-induced reduction inkernel weight/spikelets per panicle increased more than

three times with the increases of salinity. There could both primary and secondary ranchi-branches on pani-cles in sand-cultured rice plants that were treated withbe two explanations for these results. First, the compen-

sation by increasing fertility and kernel weight upon salinity at panicle initiation. In their study, secondarybranches degenerated at 0.1% NaCl (≈2.0 dS m21),the reductions in the earlier-formed yield component,

spikelet number per panicle, was significant. Second, while primary branches did not degenerate until NaClconcentration was increased to 0.5% NaCl (≈8.5 dSthe salinity effect on spikelet number per panicle was

more pronounced than that on fertility and kernel m21). Further experiments are needed to determine thethreshold value of salinity on secondary ranchi-branchesweight. The overal effects on fertility resulted from the

confounding effects of the compensation between yield in order to confirm the morphological causes of reduc-tion in spikelets on the panicle.components and the reduction due to salinity stress at

the booting and anthesis stages. It appeared that the The loss of potential spikelets has been attributed toa competition for carbohydrate supply between vegeta-salinity effect on fertility was less pronounced than that

on pollen viability (Fig. 3). Although the exact threshold tive growth and developing panicles (Murty and Murty,1982) and among spikelets within panicle (Yoshida,value of salinity effect on pollen germination was diffi-

cult to determine because of its nonlinear response to 1976). Low radiation at the early reproductive stage(Evans and De Datta, 1979; Patro and Sahu, 1986) andsalinity, the threshold values of salinity on pollen germi-

nation and fertility were both between 4.5 and 6.1 dS drought and submergence damage (Hoshikawa, 1989)may exacerbate this competition. In wheat, a completem21 (Tables 3 and 4). The close thresholds of salinity

effects on these two variables do not indicate a strong compensation between spikelets and kernels undermoderate salinity was observed on the main spike andcompensation on fertility. There might a compensation

by increasing kernel weight upon the reduction in spike- was explained by the presence of additional florets inspikelets that were usually aborted under nonstressedlets, although the multiple comparisons of mean values

were not significant (Table 4). However, a strong com- conditions, but were stimulated due to redistribution ofcarbohydrate supply upon reduction in spikelets underpensation between these two components does not seem

possible because of the enclosed hull in rice, which limits salinity (Grieve et al., 1992). In contrast, there are noadditional florets in spikelets in rice. The redistributionthe increase of grain size (Matsushima, 1957). There-

fore, the changes in the ratios with the increase of salin- of carbohydrate among spikelets upon the reduction inspikelets under salinity might partially compensate fority were mainly caused by the different sensitivity of

these yield components to salinity. It was obvious that loss of yield sink capacity by improving fertility andkernel weight among the remaining spikelets. However,plants were more sensitive to salinity at PI than to salin-

ity at later reproductive stages. It was concluded that the elongation of the culm, which occurs at the sametime as panicle differentiation and development, couldthe compensation between the successively formed yield

components was minor relative to the reduction of be highly competitive for carbohydrate supply (Moha-patra and Sahu, 1991). In M-202, shoot dry weight ofspikelets per panicle due to salinity. The highly signifi-

cant linear responses of grain weight per panicle and the main culm was not significantly different from thatof controls at moderate salinity (i.e., at or below 4.5 dSgrain weight per plant to salinity (Fig. 2) added evidence

that the compensation between these yield components m21) (Table 3). The relatively normal growth of shootsunder moderate salinity could have competed for thewas not sufficient to offset yield loss due to salinity even

at moderate salt levels. extra carbohydrate supply and constrained its distribu-tion to the remaining spikelets. The decreased harvestRice panicles consist of primary ranchi-branches, sec-

ondary branches differentiated from primary branches, index with the increase of salinity (Table 3) was consis-tent with this hypothesis. Although the mechanisms ofand flower primordia that develop into spikelets on

these branches (Hoshikawa, 1989). There is only one salinity effects on translocation of carbohydrate andassimilates among rice organs remain speculative, it isflower structure in each spikelet that will develop into a

rice kernel after fertilization and filling (Yoshida, 1981). clear that salinity reduced yield sink sites (i.e., spikeletson panicles) on main culms and restricted the intraplantThe loss of potential spikelets are due to the degenera-

tion of primary and secondary branches and flower pri- competition of yield sink organs for carbohydrate sup-ply. It is hypothesized that, under salinity, an increasemordia. The measurement of potential spikelet number

is difficult since the degeneration of branches occurs in planting density will result in an increase in the num-


chloride on germination and seedling characters of different typesber of main culms, while tillers will decrease. Harvestof rice (Oryza sativa L.). J. Agron. Crop Sci. 179:163–169.index will further decrease because of the reduced com-

Khatun, S., and T.J. Flowers. 1995. Effects of salinity on seed set inpetition of yield sink sites on main culms for carbohy- rice. Plant Cell Environ. 18:61–67.drates and shading effect at high density. Therefore, the Khatun, S., C.A. Rizzo, and T.J. Flowers. 1995. Genotypic variation

in the effect of salinity on fertility in rice. Plant Soil 173:239–250.management option of increasing planting density mightLogan, S.H., and P.B. Boyland. 1983. Calculating heat units via a sinenot be effective in dealing with yield loss under salinity.

function. J. Am. Soc. Hortic. Sci. 108:977–980.Further studies will be needed to test this hypothesis. Lutts, S., J.M. Kinet, and J. Bouharmont. 1995. Changes in plantresponse to NaCl during development of rice (Oryza sativus L.)varieties differing in salinity resistance. J. Exp. Bot. 46:1843–1852.ACKNOWLEDGMENTS

Maas, E.V., and G.J. Hoffman. 1977. Crop salt tolerance—CurrentThis research was supported by Agricultural Technology assessment. J. Irrig. Drain. Div., ASCE 103 (IR2):115–134.

Utilization and Transfer (ATUT) Project, no. 263-240. We Maas, E.V., and C.M. Grieve. 1990. Spike and leaf development insalt-stressed wheat. Crop Sci. 30:1309–1313.thank Dr. C.W. Johnson (Rice Experiment Station, Biggs, CA)

Maas, E.V., M.S. Lesch, L.E. Francois, and C.M. Grieve. 1996. Contri-for providing rice seed; Dr. C.M. Grieve, Dr. S.R. Grattan, Dr.bution of individual culms to yield of salt-stressed wheat. CropD. Mackill, and Dr. P. Siband for advising the research andSci. 36:142–149.critiquing the manuscript; P. Nash for the statistical analysis;

Matsushima, S. 1957. Analysis of development factors determiningand J. Poss, T. Donovan, and J. Draper for their excellent yield and yield prediction in lowland rice. Bull. Natl. Inst. Agric.technical assistance. Sci. Jpn. Ser. A-5 1–271.

Matsushima, S. 1970. Crop science in rice. Fuji and Publ. Co., Tokyo.Mohapatra, P.K., and S.K. Sahu. 1991. Heterogeneity of primaryREFERENCES

branch development and spikelet survival in rice panicle in relationBrewbaker, J.L., and B.H. Kwack. 1963. The essential role of calcium to assimilates of primary branches. J. Exp. Bot. 42:871–879.

ion in pollen germination and pollen tube growth. Am. J. Bot. Murty, P.S.S., and K.S. Murty. 1982. Spikelet sterility in relation to50:859–865. nitrogen and carbohydrate contents in rice. Indian J. Plant Phys-

Counce, P.A., K.A.K. Moldenhauer, and D.B. Marx. 1989. Rice yield iol. 25:40–48.and plant yield variability responses to equidistant spacing. Crop Narale, R.P., T.K. Subramangam, and R.K. Mukherjee. 1969. Influ-Sci. 29:175–179. ence of salinity on germination, vegetative growth, and grain yield

Counce, P.A., and B.R. Wells. 1990. Rice plant population density of rice (Oryza sativa var. Dular). Agron. J. 61:341–343.effect on early-season nitrogen requirement. J. Prod. Agric. Ott, L.R. 1993. An introduction to statistical methods and data analy-3:390–393. sis. 4th ed. Duxbury Press, Belmont, CA.

Cui, H., Y. Takeoka, and T. Wada. 1995. Effect of sodium chloride Patro, B., and G. Sahu. 1986. Effect of low light at different growthon the panicle and spikelet morphogenesis in rice. Jpn. J. Crop stages of the crop on sink size. Oryza 23:123–125.Sci. 64:593–600. Pearson, G.A., and L. Bernstein. 1959. Salinity effects at several

Evans, L.T., and S.K. De Datta. 1979. The relation between irradiance growth stages of rice. Agron. J. 51:654–657.and grain yield of irrigated rice in the tropics, as influenced by Sajjad, M.S. 1984. Effect of increased salt stress on yield and yieldcultivar, nitrogen fertilizer application and month of planting. Field components in rice. Pakistan J. Sci. Ind. Res. 27:292–294.Crop Res. 2:1–17. SAS Institute. 1994. SAS/STAT user’s guide. Version 6. 4th ed. SAS

Flowers, T.J., and A.R. Yeo. 1981. Variability in the resistance of Inst., Cary, NC.sodium chloride salinity within rice (Oryza sativa L.) varieties. New Scardaci, S.C., A.U. Eke, J. E. Hill, M.C. Shannon, and J.D. Rhoades.Phytol. 88:363–373. 1996. Water and soil salinity studies on California rice. Rice Publ.

Gravios, K.A., and R.S. Helms. 1992. Path analysis of rice yield and 2. Coop. Ext., Univ. of California, Colusa, CA.yield components as affected by seeding rate. Agron. J. 84:1–4. Shannon, M.C., J.D. Rhoades, J.H. Draper, S.C. Scardaci, and M.D.

Grieve, C.M., S.M. Lesch, L.E. Francois, and E.V. Maas. 1992. Analy- Spyres. 1998. Assessment of salt tolerance in rice cultivars in re-sis of main-spike yield components in salt-stressed wheat. Crop sponse to salinity problems in California. Crop Sci. 38:394–398.Sci. 32:697–703. Siband, P., J. Wey, R. Oliver, P. Letourmy, H. Manichon. 1999. Analy-

Heenan, D.P., L.G. Lewin, and D.W. McCaffery. 1988. Salinity toler- sis of the yield of two groups of tropical maize cultivars. Varietalance in rice varieties at different growth stages. Aust. J. Exp. characteristics, yield potentials, optimum densities. AgronomieAgric. 28:343–349. 19:379–394.

Hoshikawa, K. 1989. Growing rice plant—An anatomical monograph. Wu, G., L.T. Wilson, and A.M. McClung. 1998. Contribution of riceNosan Gyoson Bunka Kyokai (Nobunkyo), Tokyo, Japan. tillers to dry matter accumulation and yield. Agron. J. 90:317–323.

Kaddah, M.T. 1963. Salinity effects on growth of rice at the seedling Yoshida, S. 1976. Carbon dioxide and yield of rice. p. 211–221. Inand inflorescence stages of development. Soil Sci. 96:105–111. Climate and rice. IRRI, Los Banos, Laguna, Philippines.

Kato, T., and K. Takeda. 1996. Associations among characters related Yoshida, S., D.A. Forno, and J.H. Cock. 1971. Laboratory manual forto yield sink capacity in spaced-planted rice. Crop Sci. 36: physiological studies of rice. IRRI, Los Banos, Laguna, Philippines.1135–1139. Yoshida, S. 1981. Fundamentals of rice crop science. IRRI, Los Banos,

Laguna, Philippines.Khan, M.S.A., A. Hamid, and M.A. Karim. 1997. Effect of sodium

salinity effects on seedling growth and yield components ... · salinity effects on seedling growth...

Documents